Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler +1 authors
A method using 3D Gaussians for scene representation and optimized rendering allows high-quality, real-time novel-view synthesis at 1080p resolution.
38 篇论文 · 按点赞排序
Haohe Liu, Qiao Tian, Yi Yuan +7 authors
A unified audio generation framework using a language of audio representation, AudioMAE, and latent diffusion model achieves state-of-the-art performance across different audio types.
Peike Li, Boyu Chen, Yao Yao +3 authors
JEN-1, a diffusion model combining autoregressive and non-autoregressive training, achieves superior performance in text-to-music generation, music inpainting, and continuation, while maintaining computational efficiency.
Tianlu Wang, Ping Yu, Xiaoqing Ellen Tan +7 authors
Shepherd, a small language model tuned for critique, outperforms or ties with larger models like ChatGPT in refining language model outputs using a high-quality feedback dataset.
Yang Liu, Yuanshun Yao, Jean-Francois Ton +5 authors
This survey examines key dimensions of trustworthiness for large language models, including reliability, safety, fairness, misuse resistance, explainability, social norms adherence, and robustness, providing guidelines for evaluating and improving model alignment.
Michaël Mathieu, Sherjil Ozair, Srivatsan Srinivasan +21 authors
A benchmark called AlphaStar Unplugged is established for offline reinforcement learning using a dataset from StarCraft II, featuring unprecedented challenges and achieving high win rates with offline data.
Benjamin Spector, Chris Re
A staged speculative decoding algorithm enhances LLM inference efficiency in small-batch, on-device settings by restructuring speculative batches and adding a second decoding stage.
Xiao Liu, Hao Yu, Hanchen Zhang +19 authors
AgentBench is a multi-dimensional benchmark for evaluating LLMs as autonomous agents across various interactive environments, highlighting performance differences between commercial and open-source models.
Huichao Zhang, Bowen Chen, Hao Yang +7 authors
AvatarVerse generates high-fidelity 3D avatars from text descriptions and pose guidance using a 2D diffusion model and progressive high-resolution synthesis.
Elad Richardson, Kfir Goldberg, Yuval Alaluf +1 authors
Creative text-to-image generation leverages Diffusion Prior models to produce unique images by optimizing with adaptive constraints and creating hybrid concepts.
Wenxuan Zhou, Sheng Zhang, Yu Gu +2 authors
Targeted distillation using mission-focused instruction tuning distills ChatGPT into the smaller UniversalNER model, achieving higher NER accuracy across diverse domains than both general and state-of-the-art multi-task tuned models.
Jerry Wei, Da Huang, Yifeng Lu +2 authors
Studies reveal that language models exhibit sycophancy, altering responses to align with user views, and a synthetic-data intervention can mitigate this behavior.
Yu-Ting Lan, Kan Ren, Yansen Wang +4 authors
A pipeline named NeuroImagen uses EEG signals and a latent diffusion model to reconstruct high-resolution visual images, demonstrating superior performance in image reconstruction.
Weiran Yao, Shelby Heinecke, Juan Carlos Niebles +12 authors
A framework using policy gradient optimization to enhance large language agents by learning from environment feedback, improving their performance over time.
Weihao Yu, Zhengyuan Yang, Linjie Li +5 authors
MM-Vet evaluates large multimodal models across complicated tasks using defined vision-language capabilities and a unified LLM-based evaluator.
Jingqing Ruan, Yihong Chen, Bin Zhang +7 authors
A structured framework for LLM-based AI Agents is proposed to enhance task planning and tool usage for complex real-world applications.
Gustaf Ahdritz, Nazim Bouatta, Sachin Kadyan +7 authors
OpenProteinSet provides a large, openly available corpus of MSAs, protein homologs, and AlphaFold2 predictions, facilitating advanced training and validation for protein-focused machine learning tasks.
Ninareh Mehrabi, Palash Goyal, Christophe Dupuy +6 authors
An automatic red teaming framework uses in-context learning to identify vulnerabilities in generative models, particularly effective for text-to-image and text-to-text models like Stable Diffusion.
Roger Grosse, Juhan Bae, Cem Anil +14 authors
Influence functions, using the EK-FAC approximation, provide insights into the generalization patterns and limitations of large language models by analyzing the contribution of training examples.
Qihang Yu, Ju He, Xueqing Deng +2 authors
A single-stage framework using a shared Frozen Convolutional CLIP backbone improves open-vocabulary segmentation accuracy and efficiency compared to existing methods.
Alaa Maalouf, Ninad Jadhav, Krishna Murthy Jatavallabhula +5 authors
FAn, an open-vocabulary and multimodal robotic system, detects, tracks, and follows objects in real-time using pre-trained models and multimodal queries, and is deployable on a laptop with a lightweight GPU.
Brandon Y. Feng, Hadi Alzayer, Michael Rubinstein +2 authors
A method for 3D motion magnification that works with moving cameras and supports novel view rendering using time-varying radiance fields and the Eulerian principle.
Cliff Wong, Sheng Zheng, Yu Gu +8 authors
Large language models, like GPT-4, show promise in structuring clinical trial eligibility criteria and extracting complex matching logic, improving the efficiency and accuracy of clinical trial matching in oncology.
Tianchang Shen, Jacob Munkberg, Jon Hasselgren +7 authors
FlexiCubes improves mesh quality and fidelity in optimization by introducing flexible parameters into the Dual Marching Cubes scheme.
Hangjie Shi, Leslie Ball, Govind Thattai +39 authors
University teams competed in the SimBot Challenge to develop robot assistants that completed tasks in a simulated environment, utilizing computer vision and language models.
Sewon Min, Suchin Gururangan, Eric Wallace +3 authors
SILO is a language model that uses a parametric LM trained on a public domain corpus and a nonparametric datastore to improve performance on diverse text while mitigating legal risks.
Hu. Xuefeng, Zhang. Ke, Xia. Lu +9 authors
ReCLIP, a source-free domain adaptation method for vision-language models, reduces the average error rate of CLIP by learning a projection space and deploying cross-modality self-training with pseudo labels.
Wenqi Shao, Yutao Hu, Peng Gao +8 authors
Tiny LVLM-eHub evaluates multimodal abilities of LVLMs using ChatGPT Ensemble Evaluation, demonstrating Bard's strengths across various tasks.
Phillip Lippe, Bastiaan S. Veeling, Paris Perdikaris +2 authors
PDE-Refiner, inspired by diffusion models, provides accurate, stable, and data-efficient solutions for time-dependent PDEs by refining all frequency components.
Yilun Chen, Zhiding Yu, Yukang Chen +4 authors
A Hard Instance Probing (HIP) pipeline and its instantiation, FocalFormer3D, improve 3D object detection recall and performance through multi-stage query generation and a box-level transformer decoder.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号